I looked at Velvia 50 under a microscope today

[swilson]

<p ><a name="00Wrtm"></a><em><strong><a href="../photodb/user?user_id=449762">M M</a> <a href="../member-status-icons"></a>, Jul 15, 2010; 05:21 p.m.</strong></em></p>

 

<p><em><strong>This is as bad as it gets.Have no idea how the image from a 12 MP is as good as Velvia.You don't have to measure anything just look at them side by side.I know no fine art photographer who uses a 12 MP camera for his work.</strong></em><br>

<strong><em> </em></strong><br>

How many fine art photographers do you know who use 35mm camera for their work?</p>

 

[Komatsu]

<p>OMG!!! Now we're going to debate film vs. digital under a microscope?<br>

Whats next,,,,?!!</p>

[dave_redmann]

<p>MM, exactly what Scott said.</p>

<p>That's what we're discussing here, <em>35mm</em> film. If you want to use medium format or large format film, obviously it provides a lot more detail than the aforementioned 12-14 MP, plus (relative to 35mm) finer grain for any given output size. Obviously what Velvia equates to (insofar as you can equate it) depends on how big a piece of Velvia you use. I see B&H has it in stock in 4x5, and lists 120 as merely out of stock (by implication, still generally available). I would expect 4x5 Velvia, with the right camera/lens/technique, to be capable of outresolving even a 60 MP medium format digital back.</p>

 

[Ed_Ingold]

<p><em>How many fine art photographers do you know who use 35mm camera for their work?</em></p>

<p>There's one on the North Shore that uses an Olympus P&S. Remember! It's not the camera, it's the photographer ;-)</p>

[m_m7]

<p><strong>How many fine art photographers do you know who use 35mm camera for their work?</strong><br>

Gallen Rowell used exclusively 35mm.McCurry also most of the time.I agree in principle that the camera doesn't matter much but up to a point.Saw some tours in Yellowstone for Iphone photography where you are not allowed to bring anything else than the phone.Some guy has a bestseller on the subject(how to make great pictures with a cell phone).You make some snapshots nothing else</p>

[djcphoto]

<p>Simple test? Seriously? Do you think anyone looking at your images is going to whip out their handy microscope and analyse them? If so, you may want to question your images. Unbelievable!</p>

[thomas_lozinski]

<p>Les,<br>

The first two were underexposed. I didn't change anything else besides exposure time (then adjusted exposure in camera raw.) I'm pretty sure the purple highlights wouldn't have resulted except for the fact that the light level outside the window was much more than inside. Apparently I could have gone farther, especially if it had been an evenly lit scene. I used a Fuji S3 from 2004 (no dog when it comes to dynamic range.)</p>

<p>Frank, thanks for trying the test yourself. That was what I wanted people to do, not start a war over film vs digital. Try a lower magnification, 400x is pushing what a dry mount objective can do. Try 100x and you should clearly see the dye spots (which is different than the actual silver grains but what most people commonly refer to as grain.)</p>

<p>Jorge,<br>

Point well taken. The test only took me 1 minute and it the only reason that I did the test was that I was horribly disappointed with a canoscan 9000f flatbed scanner that I was scanning my slides on. The scans totally looked like crap so I wanted to see what the slide REALLY looked like. Obviously there is way more detail on the slide than can be picked up by low end flatbed scanners (even if it claims 9600dpi.) </p>

[ty_mickan]

<p>Thomas, resolution is one of the least important aspects of a good image.</p>

[james_symington1]

<p>Well scanned 35mm Velvia/Provia does produce more resolution than a D3/D700 but of course there are plenty of other issues that make using film an expensive pain. That said I still love it and I use much more of it now that my digital infatuation has faded rather.</p>

[mauro_franic]

<p>OP, as Edward indicated 35mm Velvia resolves 160 lpmm (this is over 4,000 true dpi) on a high contrast target. This is materially (about twice as much) higher than any 12mp camera for high contrast and fine detail.<br>

Since this is a fact (directly tested by many of us), you should use it to evaluate your methodology. Not the other way around (use your methodology to evaluate Velvia).<br>

At lower contrast the resolution goes down but not due to grain size or limitation.</p>

<p>Try to take picture through the microscope and post it for discussion.</p>

[Ed_Ingold]

<p><em>At lower contrast the resolution goes down but not due to grain size or limitation.</em></p>

<p>As you see in your microscope (or scanner), film is not really sharp. A sharply defined border (e.g., a resolution target) is spread out over several microns on the film in a roughly Gaussian fashion. As the target contrast is reduced, this spread makes it harder to distinguish edges from the background. As the frequency of the target lines increases, the spreads begin to overlap.</p>

<p>This effect is quantified in the MTF chart Fuji (and others) post for film. Notice that the contrast is fairly level at low frequencies, but begins to drop at abut 20 lp/mm. It decreases to 50% at about 50 lp/mm and only reaches 160 lp/mm at zero contrast (extrapolated). This is with an high-contrast target, of course. In practice, MTF50 is the value which determines how sharp an image looks to the eye, and at values below about 25% lines and spaces are too dim to distinguish by eye, even with magnification.</p>

[buffy_h_f.]

<blockquote>

<p><em>Of course there are individuals with better than 20/20 vision but 20/20 is the statistical standard for "perfect vision" hence it is used to calculate what the human eye can resolve.</em></p>

</blockquote>

<p>Not to quibble, but it is more accurate to call 20/20 "normal" vision rather than "perfect" vision.<em> </em></p>

<p>Buffy<em><br /></em></p>

[j_marrs]

<p>To conduct the experiment scientifically, the lens, exposure, and development of the film is critical. So if your looking at Velvia under a microscope shot with a cheap Vivitar lens and developed at Walgreens and claiming it can't out resolve digital, you're not going to have a fair comparison. If you shot it with a Leica or Rollei, exposed it properly, and had it developed at a top notch lab that keeps their chemicals fresh, then you'll have a fair comparison. Film has many more variables than digital. If the film is developed in bad chemicals, then the image will suffer. Film is a lot easier to screw up than digital. If anything in the chain of image capture and production is flawed, then the film image will suffer.<br>

From a non scientific observation, I have noticed that digital has a cleaner look than film, but film has a richer color density. The film can be cleaned up with software.</p>

[left_ayyones]

<p>Lot of <a href="http://en.wikipedia.org/wiki/Confirmation_bias">confirmation bias</a> in this thread... you guys need to relax heh...</p>

<blockquote>

<p>Confirmation bias... is a tendency for people to favor information that confirms their preconceptions or hypotheses, independently of whether they are true. As a result, people gather new evidence and recall information from memory selectively, and interpret it in a biased way. The biases appear in particular for emotionally significant issues and for established beliefs.</p>

</blockquote>

[thomas_lozinski]

<p>Here is what it looks like under the microscope. The ruler is 1 millimeter in length total divided in 100ths.<br>

I'm not set up to do microphotography so this was the best I could get aiming a point and shoot through the eyepiece. One more thing to realize is that if the dye spots/ resolution is half this means that the megapixel value would quadruple, not double. So when I say my estimation of dyespot diameter is about 10 microns, if I'm off by a factor of 10 (which could be true, this was a very quick test to get a general estimation) then the actual megapixel equivalency would be off by a factor of 100.</p><div></div>

[swilson]

<p>I don't think you are off by a factor of 10x but I think the dye clouds are closer to 6-8 microns instead of 10, IMO.</p>

[Ed_Ingold]

<p>If the dye clouds were 10 microns (8, or whatever) in diameter, Velvia could never resolve 160 lp/mm. This is known as a test of reasonableness, commonly called a "sanity check". To achieve this resolution, the <strong>maximum </strong>average diameter would be on the order of 3 microns. Perhaps you are observing something else in this photomicrogram.</p>

<p>I can't seem to find professional microphotographs or detailed information, except for a reference attributed to Kodak regarding dye clouds in color print paper at 1.75 to 4 microns in diameter. Micrographs of film to discern structure are typically taken with cross-polarized light.</p>

[jeff_livacich]

<p>Is it older Velvia? There was a problem with bubbles in the film base.</p>

[mauro_franic]

<p>Thank you Thomas.</p>

<p>Try placing a semi transparent piece of diffuse paper (like the one that separates glass slides) between the light source and the film - an as close to the film as possible. </p>

<p>Does the film look smoother this way? If so, what you might be observing are bubbles and/or imperfections on the film's protective base.</p>

<p> </p>

[swilson]

<p>So where does the 160 lp/mm come from? looking at the date sheet they only show the curve to around 65 lp/mm and at that point the MTF is already down to around 35% and dropping like a stone.<br>

<a href="http://www.fujifilm.com/products/professional_films/pdf/velvia_50_datasheet.pdf">Data sheet.</a><br>

.</p>

[swilson]

<p>This is an interesting article on grain size and resolution, it deal a lot with grain and dye clouds.<br>

<a href="http://cool.conservation-us.org/coolaic/sg/emg/library/pdf/vitale/2007-04-vitale-filmgrain_resolution.pdf">Film grain, Resolution and Fundamental Film Particles.</a></p>

[mauro_franic]

<p>Image</p><div></div>

[mauro_franic]

<p>Thomas,</p>

<p>I superimposed an actual 8000 dpi scan of velvia I took to a portion of your measuring tape TO SCALE.</p>

<p>As you can see, there are clearly 2 black lines and 1 white line resolved within the 2 marks of you 1/100 mm scale. i.e. about 160 lines per milliliter resolved.</p>

[swilson]

<p>Mauro, I am not sure what you are looking at but it looks to me more like 15 microns/lp which is 67 lp/mm. I think you are being fooled by diffraction putting in more lines then are really there. </p>

[dave_redmann]

<p><em>So where does the 160 lp/mm come from? looking at the date sheet they only show the curve to around 65 lp/mm and at that point the MTF is already down to around 35% and dropping like a stone.</em></p>

<p>Scott, look at the bottom right of page 7, where it says, "18. RESOLVING POWER, Chart Contrast 1.6:1 . . . . . 80 lines/mm, Chart Contrast 1000:1 . . . . . 160 lines/mm". Note that this says <em>lines</em> not <em>line pairs</em>. I do not know whether (1) that is an imprecise translation from Japanese usage to English usage, and they mean line pairs, or (2) that's what they really mean, lines. Obviously 80 lines/mm is 40 lp/mm and 160 lines/mm is 80 lp/mm.</p>

<p>The MTF curve in section 21 does not appear to specify the contrast. However, as previously pointed out by Edward and me on more than one occasion, by the common criterion of 50% MTF response, you're down to that at about 45 lp/mm. The curve is not drawn past about 60 lp/mm (remember, it's a log scale), at which point the MTF response is down to about 35%. Even with a perfect lens, that is not going to record real-world image detail too well. Depending on the slope of the curve past that point, maybe there is some difference in density recorded at the 160 lp/mm point, but it can't be much, and only then with very (probably artificially) high-contrast subjects.</p>

<p>Citing or testing resolution with 1000:1 contrast is very non-real-world. In old school terms, 1000:1 is the difference between Zone I and Zone XI--the extremes of what negative film might record with any real detail on either end. With transparency film like Velvia, you've got appreciable fewer zones to work with. In how many real-world pictorial subjects are many of the lines formed with that much contrast?</p>